Re: [RFC PATCH v3 19/24] x86/cet/shstk: Introduce WRUSS instruction

[Yu-cheng Yu]

On Thu, 2018-08-30 at 08:55 -0700, Andy Lutomirski wrote:
> On Thu, Aug 30, 2018 at 8:39 AM, Jann Horn <jannh@xxxxxxxxxx> wrote:
> > 
> > On Thu, Aug 30, 2018 at 4:44 PM Yu-cheng Yu <yu-cheng.yu@xxxxxxxxx
> > > wrote:
> > > 
> > > 
> > > WRUSS is a new kernel-mode instruction but writes directly
> > > to user shadow stack memory.ÂÂThis is used to construct
> > > a return address on the shadow stack for the signal
> > > handler.
> > > 
> > > This instruction can fault if the user shadow stack is
> > > invalid shadow stack memory.ÂÂIn that case, the kernel does
> > > fixup.
> > > 
> > > Signed-off-by: Yu-cheng Yu <yu-cheng.yu@xxxxxxxxx>
> > [...]
> > > 
> > > +static inline int write_user_shstk_64(unsigned long addr,
> > > unsigned long val)
> > > +{
> > > +ÂÂÂÂÂÂÂint err = 0;
> > > +
> > > +ÂÂÂÂÂÂÂasm volatile("1: wrussq %1, (%0)\n"
> > > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ"2:\n"
> > > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ_ASM_EXTABLE_HANDLE(1b, 2b,
> > > ex_handler_wruss)
> > > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ:
> > > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ: "r" (addr), "r" (val));
> > > +
> > > +ÂÂÂÂÂÂÂreturn err;
> > > +}
> > What's up with "err"? You set it to zero, and then you return it,
> > but
> > nothing can ever set it to non-zero, right?
> > 
> > > 
> > > +__visible bool ex_handler_wruss(const struct
> > > exception_table_entry *fixup,
> > > +ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂstruct pt_regs *regs, int
> > > trapnr)
> > > +{
> > > +ÂÂÂÂÂÂÂregs->ip = ex_fixup_addr(fixup);
> > > +ÂÂÂÂÂÂÂregs->ax = -1;
> > > +ÂÂÂÂÂÂÂreturn true;
> > > +}
> > And here you just write into regs->ax, but your "asm volatile"
> > doesn't
> > reserve that register. This looks wrong to me.
> > 
> > I think you probably want to add something like an explicit
> > `"+&a"(err)` output to the asm statements.
> We require asm goto support these days.ÂÂHow about using that?ÂÂYou
> won't even need a special exception handler.
> 
> Also, please change the BUG to WARN in the you-did-it-wrong 32-bit
> case.ÂÂAnd return -EFAULT.
> 
> --Andy

I will look into that.

Yu-cheng